Forest Fire Monitoring Method Based On Video Image Research And System Implementation

Forest fires often happen, not only threat the life of human but also damage the nature environment. It makes most important that find out the forest fires in early time. Most of the existing forest fire monitoring approach use manual inspection while the forest area is usually very large and difficult to find out the fires using manpower. Therefore, real-time and accurate forest fire detection methods theoretical research as well as low-cost forest fire monitoring system design has its important economic and practical significance.With the continuous development of photoelectric, smoke detectors and computer technology, the emergence of the field of forest fire monitoring composite pyrotechnic detection sensor and image type FIRE&SMOKE detection of two different development path. Due to the large range of forest fire monitoring detector type detection not only of its large-scale application of cost constraints, but optoelectronics, thermal, smoke detector detection’s performance will be greatly reduced in the complex and changing environment of the forest area. In recent years, with the promotion of the development of computer vision and pattern recognition technology, there have occurred many kinds of fire detection method based on the video image, but rarely can be well applied in the special environment of the forest area forest fire monitoring system.In this paper we focus on research of the FIRE&SMOKE monitoring based on the video image in the forest area environment, including all the aspects of the steps and methods in the process, as well as the final system implementation. The main work is as follows:(1) By the analysis and comparing of variety methods in foreground extraction, we introduced adaptive mixture Gaussian background model as the foreground detection method in the forest fire monitoring;(2) Analyze of forest fire’s static color feature and the its sharp and edge characteristics, meantime, promote the detection model based on the smoke’s area, perimeter;(3) By the introduction of the wavelet analysis method to explain the dynamic characteristics of the forest fire, reduced the false alarm rate from the motion and static feature detection;(4) Focus on reducing the cost of forest fire monitoring, and to guarantee its monitoring capabilities, we will design a complete system of forest fire monitoring, and analyze its real-time processing capabilities by using multiple types of monitoring videos.